Simulating soil C dynamics under intensive agricultural systems and climate change scenarios in the Matopiba region, Brazil

Abstract

The recent agricultural expansion in the Matopiba region, Brazil's new agricultural frontier, has raised questions about the risk of increasing soil organic carbon (SOC) loss as large areas of native vegetation (NV; i.e., Cerrado biome) have been replaced by large-scale mechanized agriculture. Although sustainable managements, such as integrated crop-livestock (ICL) systems, are considered strategic to counterbalance the SOC loss associated with land-use change (LUC) while keeping food production, little is known about their long-term effects on SOC stocks in the Matopiba region. To this end, we used the DayCent model to simulate the effects of converting the management commonly used in this region, i.e., soybean-cotton rotation under no-tillage (NT), into ICL systems with distinct levels of intensification (e.g., crop rotations: soybean-pasture and soybean-pasture-cotton; soil and crop management: grass irrigation, scarification/harrowing, and length of grass cultivation) on long term SOC dynamics. Additionally, data from two projected climate scenarios: SSP2-4.5 [greenhouse gases emissions (GHG) will not change markedly over time and global temperature will increase by 2.0 °C by 2060] and SSP5-8.5 (marked changes in GHG emissions are expected to occur resulting in an increase of 2.4 and 4.4 °C in global temperature in the middle and at the end of the century) were included in our simulations to evaluate climate change effects on SOC dynamics in this region. Based on a 50-yr-time frame simulation, we observed that SOC stocks under ICL systems were, on average, 23% and 47% higher than in the NV (36.9 Mg ha^-1) and soybean-cotton rotation under NT (30.9 Mg ha^-1), respectively. Growing grasses interlaid with crops was crucial to increase SOC stocks even when disruptive soil practices were followed. Although the irrigation of grass resulted in an early increase of SOC stocks and a higher pasture stoking rate, it did not increase SOC stocks in the long term compared to non-irrigated treatments. The SSP2-4.5 and SSP5-8.5 climate scenarios had little effects on SOC dynamics in the simulated ICL systems. However, additional SOC loss (∼0.065 Mg ha^-1 yr^-1) is predicted to occur if the current management is not improved. These findings can help guide management decisions for the Matopiba region, Brazil, to alleviate the anthropogenic pressure associated with agriculture development. More broadly, they confirm that crop-livestock integration in croplands is a successful strategy to regenerate SOC.

Keywords: Cerrado; DayCent model; Integrated crop-livestock system; Land-use change.